Secondary organic aerosol formation from evaporated biofuels: comparison to gasoline and correction for vapor wall losses

Fostering a Holistic Understanding of the Full Volatility Spectrum of Organic Compounds from Benzene Series Precursors through Mechanistic Modeling

A comprehensive understanding of the full volatility spectrum of organic oxidation products from the benzene series precursors is important to quantify the air quality and climate effects of secondary organic aerosol (SOA) and new particle formation (NPF). However, current models fail to capture the full volatility spectrum due to the absence of important reaction pathways. Here, we develop a novel unified model framework, the integrated two-dimensional volatility basis set (I2D-VBS), to simulate the full volatility spectrum of products from benzene series precursors by simultaneously representing first-generational oxidation, multigenerational aging, autoxidation, dimerization, nitrate formation, etc. The model successfully reproduces the volatility and O/C distributions of oxygenated organic molecules (OOMs) as well as the concentrations and the O/C of SOA over wide-ranging experimental conditions. In typical urban environments, autoxidation and multigenerational oxidation are the two main pathways for the formation of OOMs and SOA with similar contributions, but autoxidation contributes more to low-volatility products. NOx can reduce about two-thirds of OOMs and SOA, and most of the extremely low-volatility products compared to clean conditions, by suppressing dimerization and autoxidation. The I2D-VBS facilitates a holistic understanding of full volatility product formation, which helps fill the large gap in the predictions of organic NPF, particle growth, and SOA formation.

Modeling the Formation of Organic Compounds across Full Volatility Ranges and Their Contribution to Nanoparticle Growth in a Polluted Atmosphere

Nanoparticle growth influences atmospheric particles' climatic effects, and it is largely driven by low-volatility organic vapors. However, the magnitude and mechanism of organics' contribution to nanoparticle growth in polluted environments remain unclear because current observations and models cannot capture organics across full volatility ranges or track their formation chemistry. Here, we develop a mechanistic model that characterizes the full volatility spectrum of organic vapors and their contributions to nanoparticle growth by coupling advanced organic oxidation modeling and kinetic gas-particle partitioning. The model is applied to Nanjing, a typical polluted city, and it effectively captures the volatility distribution of low-volatility organics (with saturation vapor concentrations <0.3 μg/m3), thus accurately reproducing growth rates (GRs), with a 4.91% normalized mean bias. Simulations indicate that as particles grow from 4 to 40 nm, the relative fractions of GRs attributable to organics increase from 59 to 86%, with the remaining contribution from H2SO4 and its clusters. Aromatics contribute much to condensable organic vapors (∼37%), especially low-volatility vapors (∼61%), thus contributing the most to GRs (32-46%) as 4-40 nm particles grow. Alkanes also contribute 19-35% of GRs, while biogenic volatile organic compounds contribute minimally (<13%). Our model helps assess the climatic impacts of particles and predict future changes.

Secondary Organic Aerosol Formation from Volatile Chemical Product Emissions: Model Parameters and Contributions to Anthropogenic Aerosol

Volatile chemical products (VCP) are an increasingly important source of hydrocarbon and oxygenated volatile organic compound (OVOC) emissions to the atmosphere, and these emissions are likely to play an important role as anthropogenic precursors for secondary organic aerosol (SOA). While the SOA from VCP hydrocarbons is often accounted for in models, the formation, evolution, and properties of SOA from VCP OVOCs remain uncertain. We use environmental chamber data and a kinetic model to develop SOA parameters for 10 OVOCs representing glycols, glycol ethers, esters, oxygenated aromatics, and amines. Model simulations suggest that the SOA mass yields for these OVOCs are of the same magnitude as widely studied SOA precursors (e.g., long-chain alkanes, monoterpenes, and single-ring aromatics), and these yields exhibit a linear correlation with the carbon number of the precursor. When combined with emissions inventories for two megacities in the United States (US) and a US-wide inventory, we find that VCP VOCs react with OH to form 0.8-2.5× as much SOA, by mass, as mobile sources. Hydrocarbons (terpenes, branched and cyclic alkanes) and OVOCs (terpenoids, glycols, glycol ethers) make up 60-75 and 25-40% of the SOA arising from VCP use, respectively. This work contributes to the growing body of knowledge focused on studying VCP VOC contributions to urban air pollution.

Calculating the Enthalpy of Vaporization of Ethanol, n-Heptane, Isooctane, and a Binary Mixture of Isooctane + n-Heptane from Their Vapor Pressure Data

ConspectusA key physical property of volatile liquids is vapor pressure (VP). Volatile organic compounds (VOCs) are a classification of compounds directly associated with low boiling points, high rates of evaporation, and high flammability. The majority of chemists and chemical engineers were directly exposed to the odor of simple ethers, acetone, and toluene in the air while taking an organic chemistry laboratory course as an undergraduate student. These are just a few examples of the numerous VOCs produced by the chemical industry. When toluene is poured into a beaker from its reagent bottle, its vapors readily evaporate at ambient temperature from this open container. When the cap is securely placed back on the reagent bottle of toluene, a dynamic equilibrium develops and exists in this closed environment. This chemical concept is known as a vapor-liquid phase equilibrium. A crucial physical property of spark-ignition (SI) fuels is high volatility. In the United States, most of the vehicles traveling on the road today have SI engines. Gasoline is the fuel used to power these engines. It is a major product manufactured by the petroleum industry. This fuel is petroleum based since it is a refined product of crude oil consisting of a mixture of hydrocarbons, additives, and blending agents. Thus, gasoline is homogeneous solution of VOCs.The VP as a function of temperature of a pure VOC can readily be measured using an ebulliometer. The VP is also known in the literature as the "bubble point pressure". In this investigation, the VP as a function of temperature was acquired for the VOCs ethanol, isooctane (2,2,4-trimethylpentane), and n-heptane. The latter two VOCs are primary reference fuels components found in 87, 89, and 92 grade gasoline. Ethanol is an oxygenate additive of gasoline. The VP of a homogeneous binary mixture of isooctane and n-heptane was also acquired using the same ebulliometer and methodology. In our work, an enhanced ebulliometer was used to collect the VP data in our work. It is known as the vapor pressure acquisition system. The devices that comprise the system automatically acquire the VP data and log it into an excel spreadsheet. The data are readily transformed into information to compute the heat of vaporization (ΔH_vap). The results described in this Account compare quite favorably to the literature values. This validates our system for performing fast and reliable VP measurements.

Vapors Are Lost to Walls, Not to Particles on the Wall: Artifact-Corrected Parameters from Chamber Experiments and Implications for Global Secondary Organic Aerosol

Atmospheric models of secondary organic aerosol (OA) (SOA) typically rely on parameters derived from environmental chambers. Chambers are subject to experimental artifacts, including losses of (1) particles to the walls (PWL), (2) vapors to the particles on the wall (V2PWL), and (3) vapors to the wall directly (VWL). We present a method for deriving artifact-corrected SOA parameters and translating these to volatility basis set (VBS) parameters for use in chemical transport models (CTMs). Our process involves combining a box model that accounts for chamber artifacts (Statistical Oxidation Model with a TwO-Moment Aerosol Sectional model (SOM-TOMAS)) with a pseudo-atmospheric simulation to develop VBS parameters that are fit across a range of OA mass concentrations. We found that VWL led to the highest percentage change in chamber SOA mass yields (high NO_x: 36-680%; low NO_x: 55-250%), followed by PWL (high NO_x: 8-39%; low NO_x: 10-37%), while the effects of V2PWL are negligible. In contrast to earlier work that assumed that V2PWL was a meaningful loss pathway, we show that V2PWL is an unimportant SOA loss pathway and can be ignored when analyzing chamber data. Using our updated VBS parameters, we found that not accounting for VWL may lead surface-level OA to be underestimated by 24% (0.25 μg m^-3) as a global average or up to 130% (9.0 μg m^-3) in regions of high biogenic or anthropogenic activity. Finally, we found that accurately accounting for PWL and VWL improves model-measurement agreement for fine mode aerosol mass concentrations (PM_2.5) in the GEOS-Chem model.

Process-Level Modeling Can Simultaneously Explain Secondary Organic Aerosol Evolution in Chambers and Flow Reactors

Secondary organic aerosol (SOA) data gathered in environmental chambers (ECs) have been used extensively to develop parameters to represent SOA formation and evolution. The EC-based parameters are usually constrained to less than one day of photochemical aging but extrapolated to predict SOA aging over much longer timescales in atmospheric models. Recently, SOA has been increasingly studied in oxidation flow reactors (OFRs) over aging timescales of one to multiple days. However, these OFR data have been rarely used to validate or update the EC-based parameters. The simultaneous use of EC and OFR data is challenging because the processes relevant to SOA formation and evolution proceed over very different timescales, and both reactor types exhibit distinct experimental artifacts. In this work, we show that a kinetic SOA chemistry and microphysics model that accounts for various processes, including wall losses, aerosol phase state, heterogeneous oxidation, oligomerization, and new particle formation, can simultaneously explain SOA evolution in EC and OFR experiments, using a single consistent set of SOA parameters. With α-pinene as an example, we first developed parameters by fitting the model output to the measured SOA mass concentration and oxygen-to-carbon (O:C) ratio from an EC experiment (<1 day of aging). We then used these parameters to simulate SOA formation in OFR experiments and found that the model overestimated SOA formation (by a factor of 3-16) over photochemical ages ranging from 0.4 to 13 days, when excluding the abovementioned processes. By comprehensively accounting for these processes, the model was able to explain the observed evolution in SOA mass, composition (i.e., O:C), and size distribution in the OFR experiments. This work suggests that EC and OFR SOA data can be modeled consistently, and a synergistic use of EC and OFR data can aid in developing more refined SOA parameters for use in atmospheric models.

Heterogeneous Nucleation Drives Particle Size Segregation in Sequential Ozone and Nitrate Radical Oxidation of Catechol

Secondary organic aerosol formation via condensation of organic vapors onto existing aerosol transforms the chemical composition and size distribution of ambient aerosol, with implications for air quality and Earth's radiative balance. Gas-to-particle conversion is generally thought to occur on a continuum between equilibrium-driven partitioning of semivolatile molecules to the pre-existing mass size distribution and kinetic-driven condensation of low volatility molecules to the pre-existing surface area size distribution. However, we offer experimental evidence in contrast to this framework. When catechol is sequentially oxidized by O₃ and NO₃ in the presence of (NH₄)₂SO₄ seed particles with a single size mode, we observe a bimodal organic aerosol mass size distribution with two size modes of distinct chemical composition with nitrocatechol from NO₃ oxidation preferentially condensing onto the large end of the pre-existing size distribution (∼750 nm). A size-resolved chemistry and microphysics model reproduces the evolution of the two distinct organic aerosol size modes─heterogeneous nucleation to an independent, nitrocatechol-rich aerosol phase.

Investigation of Biofuel as a Potential Renewable Energy Source

An accelerating global energy demand, paired with the harmful environmental effects of fossil fuels, has triggered the search for alternative, renewable energy sources. Biofuels are arguably a potential renewable energy source in the transportation industry as they can be used within current infrastructures and require less technological advances than other renewable alternatives, such as electric vehicles and nuclear power. The literature suggests biofuels can negatively impact food security and production; however, this is dependent on the type of feedstock used in biofuel production. Advanced biofuels, derived from inedible biomass, are heavily favoured but require further research and development to reach their full commercial potential. Replacing fossil fuels by biofuels can substantially reduce particulate matter (PM), carbon monoxide (CO) emissions, but simultaneously increase emissions of nitrogen oxides (NOx), acetaldehyde (CH3CHO) and peroxyacetyl nitrate (PAN), resulting in debates concerning the way biofuels should be implemented. The potential biofuel blends (FT-SPK, HEFA-SPK, ATJ-SPK and HFS-SIP) and their use as an alternative to kerosene-type fuels in the aviation industry have also been assessed. Although these fuels are currently more costly than conventional aviation fuels, possible reduction in production costs has been reported as a potential solution. A preliminary study shows that i-butanol emissions (1.8 Tg/year) as a biofuel can increase ozone levels by up to 6% in the upper troposphere, highlighting a potential climate impact. However, a larger number of studies will be needed to assess the practicalities and associated cost of using the biofuel in existing vehicles, particularly in terms of identifying any modifications to existing engine infrastructure, the impact of biofuel emissions, and their chemistry on the climate and human health, to fully determine their suitability as a potential renewable energy source.

Particle Size Distribution Dynamics Can Help Constrain the Phase State of Secondary Organic Aerosol

Particle phase state is a property of atmospheric aerosols that has important implications for the formation, evolution, and gas/particle partitioning of secondary organic aerosol (SOA). In this work, we use a size-resolved chemistry and microphysics model (Statistical Oxidation Model coupled to the TwO Moment Aerosol Sectional (SOM-TOMAS)), updated to include an explicit treatment of particle phase state, to constrain the bulk diffusion coefficient (D_b) of SOA produced from α-pinene ozonolysis. By leveraging data from laboratory experiments performed in the absence of a seed and under dry conditions, we find that the D_b for SOA can be constrained ((1-7) × 10^-15 cm² s^-1 in these experiments) by simultaneously reproducing the time-varying SOA mass concentrations and the evolution of the particle size distribution. Another version of our model that used the predicted SOA composition to calculate the glass-transition temperature, viscosity, and, ultimately, D_b (∼10^-15 cm² s^-1) of the SOA was able to reproduce the mass and size distribution measurements when we included oligomer formation (oligomers accounted for about a fifth of the SOA mass). Our work highlights the potential of a size-resolved SOA model to constrain the particle phase state of SOA using historical measurements of the evolution of the particle size distribution.

Always Lost but Never Forgotten: Gas-Phase Wall Losses Are Important in All Teflon Environmental Chambers

Yields of secondary organic aerosol (SOA) formation from oxidation of volatile organic compounds are measured in laboratory chambers and then applied in regional and global models. Gas-phase losses to large Teflon-walled environmental chambers have been recently shown to reduce SOA yields. Historically, most chambers have operated in batch mode. Increasingly, however, continuous flow (CF) mode is being used, in which reactants and products are continuously introduced and exhausted from the chamber. Recent literature reports indicate a belief that SOA yields measured in CF chambers are not affected by gas-phase wall losses (GWL). Here, we use an experimentally-constrained box model to show that gas-phase wall losses impact both types of chambers when run under similar conditions. We find CF experiments do mitigate some effects of gas-phase wall losses after long (>2 days) experiment run times, but they have significant losses for typical literature experiment times of 1 day. However, this mitigation phenomenon is an experiment- and mechanism-dependent, and GWL still affects the absolute SOA yield. Finally, we show that at condensation sink values higher than the wall loss rate a lack of change in yield vs seed surface area does not necessarily indicate whether GWL affects the experiment and does not suggest the magnitude.

Oxygenated Aromatic Compounds are Important Precursors of Secondary Organic Aerosol in Biomass-Burning Emissions

Biomass burning is the largest combustion-related source of volatile organic compounds (VOCs) to the atmosphere. We describe the development of a state-of-the-science model to simulate the photochemical formation of secondary organic aerosol (SOA) from biomass-burning emissions observed in dry (RH <20%) environmental chamber experiments. The modeling is supported by (i) new oxidation chamber measurements, (ii) detailed concurrent measurements of SOA precursors in biomass-burning emissions, and (iii) development of SOA parameters for heterocyclic and oxygenated aromatic compounds based on historical chamber experiments. We find that oxygenated aromatic compounds, including phenols and methoxyphenols, account for slightly less than 60% of the SOA formed and help our model explain the variability in the organic aerosol mass (R² = 0.68) and O/C (R² = 0.69) enhancement ratios observed across 11 chamber experiments. Despite abundant emissions, heterocyclic compounds that included furans contribute to ∼20% of the total SOA. The use of pyrolysis-temperature-based or averaged emission profiles to represent SOA precursors, rather than those specific to each fire, provide similar results to within 20%. Our findings demonstrate the necessity of accounting for oxygenated aromatics from biomass-burning emissions and their SOA formation in chemical mechanisms.